Project Summary/Abstract Although it has been established that melanomas frequently have aberrant DNA methylation patterns, it is unknown if DNA methylation of primary melanoma defines biologically relevant subclasses or predicts disease- specific survival. This project's goals are to identify and characterize DNA methylation-based primary melanoma subclasses, and train, test, and validate a CpG signature prognostic for disease-specific survival in primary melanoma. Our hypothesis is that DNA methylation in primary melanoma will define subgroups (including a poor-prognostic CIMP - `CpG island methylator phenotype' subtype) and that a CpG signature will be prognostic for melanoma survival. Illumina Infinium MethylationEPIC (850K) array profiling will be performed on the same 1,000 primary melanomas to be utilized throughout this program project. These melanomas will be from AJCC stage IIA/IIB/IIC/IIIA/IIIB patients, including 500 from patients who died due to melanoma within 5 years and 500 from those who lived at least 5 years matched on stage. In Aim 1, using data from the 1,000 primary melanomas, we will identify methylation-based melanoma subclasses, characterized by 5-year survival and other outcomes, demographic factors, centrally reviewed histopathological features, a CD3/CD8/FOXP3 immune profile, PD1 and PDL1 protein expression, somatic mutations, copy number alterations, and mRNA and miRNA expression data. Further characterization will determine if the subclasses are associated with biologically relevant DNA methylation signatures that our group and others have previously produced from independent datasets. In Aim 2 using statistical modeling, we will train a primary melanoma CpG signature prognostic for death from melanoma within 5 years (the primary outcome) using 660 of the primary melanomas and test it using 340 of the primary melanomas. We will determine whether this CpG survival signature adds information to AJCC staging. As secondary outcomes, we will also develop signatures for recurrence within 5 years, sentinel lymph node (SLN) positivity, and distant recurrence after negative SLN biopsy. In Aim 3, we will technically validate DNA methylation levels of CpGs/genes in the survival signature using quantitative methylation-specific polymerase chain reaction (PCR). For select markers, we will quantitatively measure their protein expression differences using multiplexed immunohistochemistry, comparing primary melanomas from patients alive and those who died within 5 years of diagnosis and assess the function of the most prognostic targets in Project 3. Our goal is to identify and characterize DNA methylation-based primary melanoma subclasses and discover and validate a CpG signature prognostic for survival from melanoma that adds information to AJCC staging. We will initiate application of the signature to a clinically viable assay prognostic for survival. This project should allow identification of melanoma patients expected to have worse survival and who could benefit from closer follow-up and adjuvant treatments. Our results should also indicate which targets predictive of treatment responses are present in poor prognostic melanomas.